1 /*
2  * Copyright © 2012-2018 Inria.  All rights reserved.
3  * See COPYING in top-level directory.
4  *
5  * Modifications after import:
6  * - removed all #if
7  * - updated prototypes
8  * - updated #include
9  */
10 
11 /* include hwloc's config before anything else
12  * so that extensions and features are properly enabled
13  */
14 #include "private/private.h"
15 
16 /*	$OpenBSD: base64.c,v 1.5 2006/10/21 09:55:03 otto Exp $	*/
17 
18 /*
19  * Copyright (c) 1996 by Internet Software Consortium.
20  *
21  * Permission to use, copy, modify, and distribute this software for any
22  * purpose with or without fee is hereby granted, provided that the above
23  * copyright notice and this permission notice appear in all copies.
24  *
25  * THE SOFTWARE IS PROVIDED "AS IS" AND INTERNET SOFTWARE CONSORTIUM DISCLAIMS
26  * ALL WARRANTIES WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES
27  * OF MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL INTERNET SOFTWARE
28  * CONSORTIUM BE LIABLE FOR ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL
29  * DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR
30  * PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS
31  * ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS
32  * SOFTWARE.
33  */
34 
35 /*
36  * Portions Copyright (c) 1995 by International Business Machines, Inc.
37  *
38  * International Business Machines, Inc. (hereinafter called IBM) grants
39  * permission under its copyrights to use, copy, modify, and distribute this
40  * Software with or without fee, provided that the above copyright notice and
41  * all paragraphs of this notice appear in all copies, and that the name of IBM
42  * not be used in connection with the marketing of any product incorporating
43  * the Software or modifications thereof, without specific, written prior
44  * permission.
45  *
46  * To the extent it has a right to do so, IBM grants an immunity from suit
47  * under its patents, if any, for the use, sale or manufacture of products to
48  * the extent that such products are used for performing Domain Name System
49  * dynamic updates in TCP/IP networks by means of the Software.  No immunity is
50  * granted for any product per se or for any other function of any product.
51  *
52  * THE SOFTWARE IS PROVIDED "AS IS", AND IBM DISCLAIMS ALL WARRANTIES,
53  * INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A
54  * PARTICULAR PURPOSE.  IN NO EVENT SHALL IBM BE LIABLE FOR ANY SPECIAL,
55  * DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER ARISING
56  * OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE, EVEN
57  * IF IBM IS APPRISED OF THE POSSIBILITY OF SUCH DAMAGES.
58  */
59 
60 /* OPENBSD ORIGINAL: lib/libc/net/base64.c */
61 
62 static const char Base64[] =
63 	"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
64 static const char Pad64 = '=';
65 
66 /* (From RFC1521 and draft-ietf-dnssec-secext-03.txt)
67    The following encoding technique is taken from RFC 1521 by Borenstein
68    and Freed.  It is reproduced here in a slightly edited form for
69    convenience.
70 
71    A 65-character subset of US-ASCII is used, enabling 6 bits to be
72    represented per printable character. (The extra 65th character, "=",
73    is used to signify a special processing function.)
74 
75    The encoding process represents 24-bit groups of input bits as output
76    strings of 4 encoded characters. Proceeding from left to right, a
77    24-bit input group is formed by concatenating 3 8-bit input groups.
78    These 24 bits are then treated as 4 concatenated 6-bit groups, each
79    of which is translated into a single digit in the base64 alphabet.
80 
81    Each 6-bit group is used as an index into an array of 64 printable
82    characters. The character referenced by the index is placed in the
83    output string.
84 
85                          Table 1: The Base64 Alphabet
86 
87       Value Encoding  Value Encoding  Value Encoding  Value Encoding
88           0 A            17 R            34 i            51 z
89           1 B            18 S            35 j            52 0
90           2 C            19 T            36 k            53 1
91           3 D            20 U            37 l            54 2
92           4 E            21 V            38 m            55 3
93           5 F            22 W            39 n            56 4
94           6 G            23 X            40 o            57 5
95           7 H            24 Y            41 p            58 6
96           8 I            25 Z            42 q            59 7
97           9 J            26 a            43 r            60 8
98          10 K            27 b            44 s            61 9
99          11 L            28 c            45 t            62 +
100          12 M            29 d            46 u            63 /
101          13 N            30 e            47 v
102          14 O            31 f            48 w         (pad) =
103          15 P            32 g            49 x
104          16 Q            33 h            50 y
105 
106    Special processing is performed if fewer than 24 bits are available
107    at the end of the data being encoded.  A full encoding quantum is
108    always completed at the end of a quantity.  When fewer than 24 input
109    bits are available in an input group, zero bits are added (on the
110    right) to form an integral number of 6-bit groups.  Padding at the
111    end of the data is performed using the '=' character.
112 
113    Since all base64 input is an integral number of octets, only the
114          -------------------------------------------------
115    following cases can arise:
116 
117        (1) the final quantum of encoding input is an integral
118            multiple of 24 bits; here, the final unit of encoded
119 	   output will be an integral multiple of 4 characters
120 	   with no "=" padding,
121        (2) the final quantum of encoding input is exactly 8 bits;
122            here, the final unit of encoded output will be two
123 	   characters followed by two "=" padding characters, or
124        (3) the final quantum of encoding input is exactly 16 bits;
125            here, the final unit of encoded output will be three
126 	   characters followed by one "=" padding character.
127    */
128 
129 #include <stdlib.h>
130 #include <string.h>
131 #include <ctype.h>
132 
133 int
hwloc_encode_to_base64(const char * src,size_t srclength,char * target,size_t targsize)134 hwloc_encode_to_base64(const char *src, size_t srclength, char *target, size_t targsize)
135 {
136 	size_t datalength = 0;
137 	unsigned char input[3];
138 	unsigned char output[4];
139 	unsigned int i;
140 
141 	while (2 < srclength) {
142 		input[0] = *src++;
143 		input[1] = *src++;
144 		input[2] = *src++;
145 		srclength -= 3;
146 
147 		output[0] = input[0] >> 2;
148 		output[1] = ((input[0] & 0x03) << 4) + (input[1] >> 4);
149 		output[2] = ((input[1] & 0x0f) << 2) + (input[2] >> 6);
150 		output[3] = input[2] & 0x3f;
151 
152 		if (datalength + 4 > targsize)
153 			return (-1);
154 		target[datalength++] = Base64[output[0]];
155 		target[datalength++] = Base64[output[1]];
156 		target[datalength++] = Base64[output[2]];
157 		target[datalength++] = Base64[output[3]];
158 	}
159 
160 	/* Now we worry about padding. */
161 	if (0 != srclength) {
162 		/* Get what's left. */
163 		input[0] = input[1] = input[2] = '\0';
164 		for (i = 0; i < srclength; i++)
165 			input[i] = *src++;
166 
167 		output[0] = input[0] >> 2;
168 		output[1] = ((input[0] & 0x03) << 4) + (input[1] >> 4);
169 		output[2] = ((input[1] & 0x0f) << 2) + (input[2] >> 6);
170 
171 		if (datalength + 4 > targsize)
172 			return (-1);
173 		target[datalength++] = Base64[output[0]];
174 		target[datalength++] = Base64[output[1]];
175 		if (srclength == 1)
176 			target[datalength++] = Pad64;
177 		else
178 			target[datalength++] = Base64[output[2]];
179 		target[datalength++] = Pad64;
180 	}
181 	if (datalength >= targsize)
182 		return (-1);
183 	target[datalength] = '\0';	/* Returned value doesn't count \0. */
184 	return (int)(datalength);
185 }
186 
187 /* skips all whitespace anywhere.
188    converts characters, four at a time, starting at (or after)
189    src from base - 64 numbers into three 8 bit bytes in the target area.
190    it returns the number of data bytes stored at the target, or -1 on error.
191  */
192 
193 int
hwloc_decode_from_base64(char const * src,char * target,size_t targsize)194 hwloc_decode_from_base64(char const *src, char *target, size_t targsize)
195 {
196 	unsigned int tarindex, state;
197 	int ch;
198 	char *pos;
199 
200 	state = 0;
201 	tarindex = 0;
202 
203 	while ((ch = *src++) != '\0') {
204 		if (isspace(ch))	/* Skip whitespace anywhere. */
205 			continue;
206 
207 		if (ch == Pad64)
208 			break;
209 
210 		pos = strchr(Base64, ch);
211 		if (pos == 0) 		/* A non-base64 character. */
212 			return (-1);
213 
214 		switch (state) {
215 		case 0:
216 			if (target) {
217 				if (tarindex >= targsize)
218 					return (-1);
219 				target[tarindex] = (char)(pos - Base64) << 2;
220 			}
221 			state = 1;
222 			break;
223 		case 1:
224 			if (target) {
225 				if (tarindex + 1 >= targsize)
226 					return (-1);
227 				target[tarindex]   |=  (pos - Base64) >> 4;
228 				target[tarindex+1]  = ((pos - Base64) & 0x0f)
229 							<< 4 ;
230 			}
231 			tarindex++;
232 			state = 2;
233 			break;
234 		case 2:
235 			if (target) {
236 				if (tarindex + 1 >= targsize)
237 					return (-1);
238 				target[tarindex]   |=  (pos - Base64) >> 2;
239 				target[tarindex+1]  = ((pos - Base64) & 0x03)
240 							<< 6;
241 			}
242 			tarindex++;
243 			state = 3;
244 			break;
245 		case 3:
246 			if (target) {
247 				if (tarindex >= targsize)
248 					return (-1);
249 				target[tarindex] |= (pos - Base64);
250 			}
251 			tarindex++;
252 			state = 0;
253 			break;
254 		}
255 	}
256 
257 	/*
258 	 * We are done decoding Base-64 chars.  Let's see if we ended
259 	 * on a byte boundary, and/or with erroneous trailing characters.
260 	 */
261 
262 	if (ch == Pad64) {		/* We got a pad char. */
263 		ch = *src++;		/* Skip it, get next. */
264 		switch (state) {
265 		case 0:		/* Invalid = in first position */
266 		case 1:		/* Invalid = in second position */
267 			return (-1);
268 
269 		case 2:		/* Valid, means one byte of info */
270 			/* Skip any number of spaces. */
271 			for (; ch != '\0'; ch = *src++)
272 				if (!isspace(ch))
273 					break;
274 			/* Make sure there is another trailing = sign. */
275 			if (ch != Pad64)
276 				return (-1);
277 			ch = *src++;		/* Skip the = */
278 			/* Fall through to "single trailing =" case. */
279 			/* FALLTHROUGH */
280 
281 		case 3:		/* Valid, means two bytes of info */
282 			/*
283 			 * We know this char is an =.  Is there anything but
284 			 * whitespace after it?
285 			 */
286 			for (; ch != '\0'; ch = *src++)
287 				if (!isspace(ch))
288 					return (-1);
289 
290 			/*
291 			 * Now make sure for cases 2 and 3 that the "extra"
292 			 * bits that slopped past the last full byte were
293 			 * zeros.  If we don't check them, they become a
294 			 * subliminal channel.
295 			 */
296 			if (target && target[tarindex] != 0)
297 				return (-1);
298 		}
299 	} else {
300 		/*
301 		 * We ended by seeing the end of the string.  Make sure we
302 		 * have no partial bytes lying around.
303 		 */
304 		if (state != 0)
305 			return (-1);
306 	}
307 
308 	return (tarindex);
309 }
310